-
Notifications
You must be signed in to change notification settings - Fork 1
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
1 parent
6023ea4
commit e9a8d71
Showing
2 changed files
with
205 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,205 @@ | ||
| --- | ||
| slug: 9月22日内容汇总 | ||
| title: 9月22日内容汇总 | ||
| authors: [garfield] | ||
| tags: [] | ||
| --- | ||
|
|
||
|  | ||
|
|
||
| 封面图：Safe by construction - Roberto Clapis | ||
|
|
||
| ## 🌟 AI 相关 | ||
|
|
||
| :::tip 技术资讯 | ||
|
|
||
| - [🥳来正式介绍下 LobeChat v1.19 版本的新功能](https://x.com/arvin17x/status/1837097802142470587) | ||
| - 支持 Artifacts | ||
| - 『发现』 页面全面焕新 | ||
| - 支持 GitHub Models 模型服务商 | ||
| - [《Hands-On Large Language Models》在本周读完了，这是今年唯二的我读过的关于大模型的最棒的书，另一本是《Build a Large Language Model (From Scratch)》](https://x.com/chenqing663/status/1837511826902601970) | ||
| - [永不贬值的16大资产](https://x.com/didengshengwu/status/1837110305102659668) | ||
| - [Really 🤯. Qwen 2.5 is outranking GPT-4o and o1-preview on Livebench coding category.](https://x.com/rohanpaul_ai/status/1837557221288186260) | ||
| - [Prompt chaining is a really useful technique to build more robust LLM applications. ](https://x.com/omarsar0/status/1836854828758356199) | ||
| - [RAG 技术精髓：检索、增强与生成的完美融合](https://x.com/shao__meng/status/1837056458292351405) | ||
| - [撞了100次南墙才明白的80条道理！](https://x.com/caijingshujuku/status/1836980683878580604) | ||
| - [现在是开始卷 AI 编辑器了吗？昨天看到一个类似 Cursor 的开源编辑器 - Void](https://x.com/vikingmute/status/1836579797558005868) | ||
| - [延迟 200ms，仅 ~7.6B，基于 Rust 实现推理库！🔥 支持在 Macbook Air 上端侧运行！端到端的语音会话式 AI - Moshi 正式宣布开源 ⚡️](https://x.com/tuturetom/status/1836940136929714385) | ||
| - [四张图 看懂本次降息的机遇与发展！！](https://x.com/didengshengwu/status/1836725850034876682) | ||
| - [It’s time 🍓. Today, we’re thrilled to open up early access to OpenAI o1- preview and o1-mini for GitHub Copilot Chat in VSCode. And there’s more: it’s also available in our GitHub models catalog, today!](https://x.com/mariorod1/status/1836855226864742827) | ||
| - [Very Powerful but very simple Prompting technique. Simply ask the LLM to re-read the question - and this significantly boosts LLM reasoning across diverse tasks and model types. 💡](https://x.com/rohanpaul_ai/status/1836890159314522445) | ||
| - [RAG 上下文检索最佳技术实践发布！⚡️ @AnthropicAI 将价格打到 1.02 美元/百万 token，并降低高达 67% 的错误率🔥](https://x.com/tuturetom/status/1836944896101601311) | ||
| - [拿捏人性 最硬核的手段](https://github.com/user-attachments/assets/be701bd3-934d-4113-9e5f-513f87db249e) | ||
| - [Advanced RAG (Notebooks) 教程涵盖从基础到高级 RAG 技术的各个方面，包括多查询检索、自反射 RAG、Agentic RAG 等多种高级技术](https://x.com/shao__meng/status/1836712960196305163) | ||
| - [微软推出了将web + work + Page结合在一起的新型办公工具：Copilot Pages，工作协作起来方便很多](https://x.com/aigclink/status/1835931927658029442) | ||
| - [这是我见过最简洁的博客&备忘录 GitHub 一键登录，编写并同步，数据直接存储到 GitHub 仓库。👍](https://github.com/mazzzystar/tinymind) | ||
| - [最新Qwen2.5刚刚发布了，这次一口气开源了一大串模型！有语言模型 Qwen2.5，编程模型Qwen2.5-Coder、数学模型 Qwen2.5-Math](https://x.com/aigclink/status/1836545580031746278) | ||
| - [Meta 官方推荐 Together AI - LlamaCoder @AIatMeta @togethercompute](https://x.com/shao__meng/status/1836556061383496000) | ||
| - [30条"过来人"的人生智慧](https://x.com/didengshengwu/status/1836581565251653974) | ||
| - [人生22大忌讳](https://x.com/didengshengwu/status/1836383055277363710) | ||
| - [Google DeepMind 研究科学家 @agarwl_ 分享的 50+ 页的 PPT 《使用 RL 和验证器改进 LLM 推理》也太好了🔥](https://x.com/tuturetom/status/1836449582747939260) | ||
| - [Welcome to the party of Qwen2.5 foundation models! This time, we have the biggest release ever in the history of Qwen.](https://x.com/Alibaba_Qwen/status/1836449414220779584) | ||
| - ["操纵心理学"的七种常见技巧及应对策略](https://x.com/0427SMtieshou/status/1836200967773225307) | ||
| - [能量高的人通常都是柔软的](https://github.com/user-attachments/assets/7945a445-cc20-4141-a643-caecd63ecd2f) | ||
| - [275+ 插画，12+ 章节，吴恩达力荐！⚡️完美呈现 LLM 工业级应用，提供完整可运行的代码和 Paper 引用🤯🤯，这本 《实践大语言模型》也太好了🔥](https://x.com/tuturetom/status/1835880356475724141) | ||
| - [书籍+代码：动手实践 LLM ](https://x.com/shao__meng/status/1835836908360093970) | ||
| - [Chain of Thought (CoT) enables transformers to execute serial computations, expanding their problem-solving capabilities beyond parallel-only limitations.](https://x.com/rohanpaul_ai/status/1835868073666207863) | ||
| - [40岁来临前,要懂得人性的12铁律](https://x.com/didengshengwu/status/1836245252145057977) | ||
| - [到底选择什么方式挣钱？各种方式的本质核心是什么？有什么优劣？哪种更适合自己？信息爆炸，我们看了太多无用的东西。简简单单一张图，看一下，是不是让你豁然开朗？](https://x.com/0427SMtieshou/status/1836925744334635385) | ||
| - [OpenAI CEO @sama 宣布自家 AI 目前已经达到 Goal 3，即：⚡️🤯🤯](https://x.com/tuturetom/status/1836225295755677737) | ||
| - [Multimodal RAG over Product Manuals 📖🎨](https://x.com/llama_index/status/1835813824005583313) | ||
| - [破解了，现在 o1 两个版本的系统提示词极其简单： You are ChatGPT, a large language model trained by OpenAI, based on the GPT-4 architecture. Knowledge cutoff: 2023-10 Current date: 2024-09-17](https://x.com/dotey/status/1836130348247454018) | ||
| - [Hey, Qwen2.5-72B-Instruct is now on LMSys Chatbot Arena, @lmsysorg](https://x.com/Alibaba_Qwen/status/1836063387085934909) | ||
| - [A cool Github repo collecting LLM papers, blogs, and projects, with a focus on OpenAI o1 and reasoning techniques.](https://x.com/rohanpaul_ai/status/1835471217014255941) | ||
|
|
||
| ::: | ||
|
|
||
| [最近大厂推出的Prompt Cache到底是个啥](https://mp.weixin.qq.com/s/haFTftqsmiD7_bnXpAhWrA) | ||
|
|
||
| [现代LLM基本技术整理](https://mp.weixin.qq.com/s/d1EDexcu7I57km-Fz5uzHA) | ||
|
|
||
| [实现一个简洁的代码模型评测框架（以Qwen2.5-coder 评测Humaneval为例）](https://mp.weixin.qq.com/s/pvtrtVrjrl27Ilugf3VwkQ) | ||
|
|
||
| [首个Mamba+Transformer混合架构多模态大模型来了，实现单卡千图推理](https://mp.weixin.qq.com/s/ipfx6qaxeQlkILEVACabvA) | ||
|
|
||
| [OpenAI前研究者发布提示词工程框架ell，升级版LangChain，支持版本控制和多模态](https://mp.weixin.qq.com/s/LaNbu4bVrWLG3ueopFTj5g) | ||
|
|
||
| [SGLang v0.3发布！并向你发来一封邀请函](https://mp.weixin.qq.com/s/_yTe8qzTMVD3yHC8uGLQOQ) | ||
|
|
||
| [OpenAI o1的合成数据与推理搜索是怎么实现的](https://mp.weixin.qq.com/s/tV6YSLsREhGgWBT_GN2iCg) | ||
|
|
||
| [o1带火的CoT到底行不行？新论文引发了论战](https://mp.weixin.qq.com/s/_v15-UYpv300XIlhQ-54Vg) | ||
|
|
||
| [社区供稿 | Reader-LM: 将原始 HTML 转换为干净 Markdown 的小型语言模型](https://mp.weixin.qq.com/s/1run-59DhhyvsZnR7fQF5A) | ||
|
|
||
| [Anthropic分享RAG最佳实践：Contextual Retrieval！](https://mp.weixin.qq.com/s/LEtk3iFgSsfbKuD7HhRRew) | ||
|
|
||
| [大模型Agent，如何利用历史经验自我进化](https://mp.weixin.qq.com/s/0UJdTK9MSQjXNIC78oZZgQ) | ||
|
|
||
| [Agent+RAG+大纲驱动，AI创作新风暴](https://mp.weixin.qq.com/s/qvAG_wy8lFjbmrxOlTve7g) | ||
|
|
||
| [CoT能让模型推理能力无上限？田渊栋、LeCun下场反对：两层MLP还能模拟全世界呢](https://mp.weixin.qq.com/s/wi1jvQg47O078Xk83UpYmA) | ||
|
|
||
| [Seed-Music 音乐大模型正式发布！生成编辑两开花，十种创作任务，满足多样化需求](https://mp.weixin.qq.com/s/jNtNIbN0VCLVi9JN8UisNA) | ||
|
|
||
| [TC-RAG: 图灵完备的检索增强](https://mp.weixin.qq.com/s/K7DFiVAocR1nTCSqMxOuEA) | ||
|
|
||
| [Jina又开源向量V3，8K长度，卷疯了！](https://mp.weixin.qq.com/s/i7jradq7Xg3Dl1KA_NICpQ) | ||
|
|
||
| [Qwen2.5系列模型开源，你值得拥有！！！](https://mp.weixin.qq.com/s/icamjnmSy6Mq-SMd1LxEfQ) | ||
|
|
||
| [Qwen2.5 全链路模型体验、下载、推理、微调、部署实战！](https://mp.weixin.qq.com/s/Gn8kV04e_Y_BmXdxSMBiSQ) | ||
|
|
||
| [Qwen2.5系列开源来袭！ 冲！](https://mp.weixin.qq.com/s/k6ao152XYL1J3PpcezXZ4Q) | ||
|
|
||
| [g1：基于 Llama，用提示工程实现类似 o1 的深度推理](https://mp.weixin.qq.com/s/22ImXInUkGJiE3VvJ3BEQA) | ||
|
|
||
| 🌟 [30 天 52% 回报：GPT-4o 量化交易机器人](https://mp.weixin.qq.com/s/nRSTqguLVK7qTcLUg5Lv8w) | ||
|
|
||
| 🌟 [开源！我用Qwen2VL实现了一个多模态RAG](https://mp.weixin.qq.com/s/Ks33a1nytFfQ32mGbNob6g) | ||
|
|
||
| [一个包含15种大模型Agent技巧的项目开源](https://mp.weixin.qq.com/s/_0-xyd7zX1W5tlaGAdT_PA) | ||
|
|
||
| [DsRAG：新开源的高级RAG分块技巧！](https://mp.weixin.qq.com/s/QUM_PTGFtbJkyl9pqjmLdA) | ||
|
|
||
| [OpenAI o1的开源平替版self-replay RL来了！](https://mp.weixin.qq.com/s/KlLU3eHsFn0qo0N8nmqK9g) | ||
|
|
||
| [HippoRAG:模拟大脑记忆思索机制的 RAG](https://mp.weixin.qq.com/s/LKVMAZmXeEbyklsIZGjqqQ) | ||
|
|
||
| [SGLang：地表最强LLM推理引擎安装经验分享，以及为何频繁掉卡](https://mp.weixin.qq.com/s/EAwz6-6JzZa77RzYobyMTw) | ||
|
|
||
| [微软Azure AI检索分享：通过混合检索和语义重排超越向量搜索](https://mp.weixin.qq.com/s/quuvzTw40VDuno9JxJKpjw) | ||
|
|
||
| [人人都能看懂的DPO数学原理](https://mp.weixin.qq.com/s/WjVZ3I5zytpf5md53KKuog) | ||
|
|
||
| [简单图解一下线性注意力机制](https://mp.weixin.qq.com/s/7P-kwV9yaOsbsC2OZaJL1w) | ||
|
|
||
| [OpenAI o1背后的技术：LLM的快思考与慢思考路线之MCTS](https://mp.weixin.qq.com/s/XBZzCOB6Q0Y9IClmeAauDw) | ||
|
|
||
| [技术上，如何复现 o1](https://mp.weixin.qq.com/s/An54ot8C8OQiywhysApfBQ) | ||
|
|
||
| 🌟 [RAG高级优化：一文看尽query的转换之路](https://mp.weixin.qq.com/s/UZV4ztL0j5k2qPV8U_wZVg) | ||
|
|
||
| [别以为Cursor生成不了屎山代码，用PLANSEARCH给你更多计划，性能提升到77%| CursorAI最新](https://mp.weixin.qq.com/s/iQ_mOxFEJOJICwfDQTIu5w) | ||
|
|
||
| [别吹了，OpenAI-o1绝没吹的那么神，用斯坦福的这条Prompt，一样实现说话之前先思考](https://mp.weixin.qq.com/s/os6pOy723hxA4I6Apw8-1g) | ||
|
|
||
| [长文 | 探索基于RL的新LLM scaling范式](https://mp.weixin.qq.com/s/6XwrSiOjKAD9oXhmDx8_GA) | ||
|
|
||
| ## ⭐️ Go & 云原生 & Rust 相关 | ||
|
|
||
| :::tip 技术资讯 | ||
|
|
||
| - [Delve into data types and casting in Rust. Learn why they matter, see clear syntax examples, and follow my learning journey.](https://x.com/the_Rust_Dev/status/1837402636477243592) | ||
| - [Explore how to construct a simple Rust-based neural network from scratch. Start building, implement core components, and train on a dataset today.](https://x.com/the_Rust_Dev/status/1837333936797589927) | ||
| - [用自己的工具迁移数据，Apple Store 店员都看傻了，说明年可以推广一下这个 app 🤣](https://x.com/unixzii/status/1837376659932041338) | ||
| - 项目地址：https://github.com/unixzii/ibackupextractor | ||
| - [这个问题变更只有一行，但是查了我很久的时间，反复在 Go 和 Glibc 的源码中横跳。顺便还去复习了 Linker 的很多知识](https://www.manjusaka.blog/posts/2024/09/18/a-live-debug-ecapture-gh-604/) | ||
| - [Bare metal to production ready in mins; your own fly server on your VPS. #golang based](https://x.com/golangch/status/1836850979708244288) | ||
| - [Golang vs Ruby on pointers: understanding how data gets referenced can lead to cleaner code and efficient memory use. Dive into the details. ](https://x.com/TheGoDev/status/1836727188445384709) | ||
| - [A performance monitoring library for Go apps, offering real-time insights into service-level and function-level metrics #golang](https://x.com/golangch/status/1836626713070952613) | ||
| - [GORM multi-tenant applications, streamlining tenant management and model migrations. Abstracting multitenancy complexities through a unified, database-agnostic API compatible with GORM.](https://x.com/golangch/status/1836391702149353836) | ||
| - [Learn how to implement a simple RPC interface in Go using the net/rpc package. Understand the essentials of Remote Procedure Calls.](https://x.com/TheGoDev/status/1836569147817259378) | ||
| - [Discover the power of Go's empty struct. How does zero-sized data carry a purpose, and use it to save memory efficiently.](https://x.com/TheGoDev/status/1836437276210958721) | ||
| - [Dive into the difference between unbuffered and buffered channels in Go—from behavior nuances to effective usage scenarios!](https://x.com/TheGoDev/status/1836047708865372303) | ||
| - [“Interning in Go“ #golang](https://medium.com/google-cloud/interning-in-go-4319ea635002) | ||
| - [Dive into practical Golang logging tips: use Go's log package, explore third-party libraries, structured logging, and performance insights.](https://x.com/TheGoDev/status/1835685321888997766) | ||
| - [《13 年后，我如何用 Go 写 HTTP 服务》](https://grafana.com/blog/2024/02/09/how-i-write-http-services-in-go-after-13-years/) | ||
|
|
||
| ::: | ||
|
|
||
| [GWS 项目的秘密武器：高性能 bytes.Buffer 池设计详解](https://mp.weixin.qq.com/s/rZ3Ec4ESZvV5-hipagvs_Q) | ||
|
|
||
| [万字解析 golang netpoll 底层原理](https://mp.weixin.qq.com/s/_FTvpvLIWfYzgNhOJgKypA) | ||
|
|
||
| [Go 入门指南：9.1. 标准库概述](https://mp.weixin.qq.com/s/4UInvgN9DjrTxF9GfuK8Kw) | ||
|
|
||
| [htmx：Gopher走向全栈的完美搭档](https://mp.weixin.qq.com/s/081HOITO0_PQG3dsKfwBCw) | ||
|
|
||
| [Go语言轻量级内存数据库，基于Redis哈希表！](https://mp.weixin.qq.com/s/AhAPas92lVrW46-3SrM8og) | ||
|
|
||
| [Go 入门指南：9.2. regexp 包](https://mp.weixin.qq.com/s/UpVik6U2tO_xfFXj3u4uaw) | ||
|
|
||
| [Golang Goroutine 切换的耗时分析](https://mp.weixin.qq.com/s/zNcbMVKe86Ay7jynN97JPw) | ||
|
|
||
| [Go 入门指南：8.5. map 的排序](https://mp.weixin.qq.com/s/-hwRAKy5OEXWcNk6dF0pPQ) | ||
|
|
||
| [Go 语言中的 JSON 库 V2 版本](https://mp.weixin.qq.com/s/NFF5Px7TaOhM3DNkPLD3fg) | ||
|
|
||
| [Go unique包：突破字符串局限的通用值Interning技术实现](https://mp.weixin.qq.com/s/SiKFOZvaqz5Gwjl6OgbujQ) | ||
|
|
||
| [构建由大型语言模型（LLM）驱动的 Go 应用程序](https://mp.weixin.qq.com/s/W5OpauBtdhsUu6AEqeP53w) | ||
|
|
||
| [为什么Rust中的Mutex设计成如此这般](https://mp.weixin.qq.com/s/Pnj8w2YDV9biIDsSP95nkQ) | ||
|
|
||
| [Go必知必会：测试--编写可靠代码的实践指南](https://mp.weixin.qq.com/s/wD8B8UUdB37y4mAYPnGbjA) | ||
|
|
||
| [Go 入门指南：8.3. for-range 的配套用法](https://mp.weixin.qq.com/s/gsR-Qr7r6wwz7enckse2mA) | ||
|
|
||
| [eBPF 开发实战 - 03 fentry/fexit入门](https://mp.weixin.qq.com/s/vjSzxRnK5iXj9X7CN3WUjA) | ||
|
|
||
| [Rust 的错误处理真的完美吗](https://mp.weixin.qq.com/s/mhHJQnBW2-fdMneSWpnYLw) | ||
|
|
||
| [Go必知必会：部署（下）通过Shell脚本一键部署Go项目到服务器](https://mp.weixin.qq.com/s/qNKP5DSjF3-UMSOvfY-msA) | ||
|
|
||
| [Go 入门指南：8.2. 测试键值对是否存在及删除元素](https://mp.weixin.qq.com/s/FBNNOKDZ4ellF4hRFx24RQ) | ||
|
|
||
| ## 📒 后端相关 | ||
|
|
||
| [MySQL核心之InnoDB Buffer Pool：图文透视](https://mp.weixin.qq.com/s/eFPmzgrirpbg60Kw85mp_g) | ||
|
|
||
| [美团面试：binlog、redolog、undo log底层原理是啥？分别实现ACID哪个特性？（尼恩图解，史上最全）](https://mp.weixin.qq.com/s/p1eHHIVPetADdrWe__Hdbw) | ||
|
|
||
| [快手二面: 千万级别数据的订单表,给订单状态加索引有用嘛](https://mp.weixin.qq.com/s/HBgtKmMCuC3hgdvHgcqIvA) | ||
|
|
||
| [我司使用了两年的高效日志打印工具，非常好用！](https://mp.weixin.qq.com/s/xyOyI4PdwgH8mr-mhVzYIw) | ||
|
|
||
| [快速基于 ClickHouse + Grafana 搭建可观测性解决方案 - 日志篇(ClickHouse 官方博客)](https://mp.weixin.qq.com/s/KMv4IbkctLJgqAJm6xWWvQ) | ||
|
|
||
| [Binance 如何使用 Quickwit 构建 100PB 日志服务(Quickwit 博客)](https://mp.weixin.qq.com/s/ubfdwJ5c3G_LjblG50YUaw) | ||
|
|
||
| ## 📒 前端相关 | ||
|
|
||
| [ChatGPT 从 Next.js 迁移到 Remix](https://mp.weixin.qq.com/s/pCP3wKoj3nb97sz6lZ_IdA) | ||
|
|
||
| [Cursor 为什么这么牛逼，以每秒 1000 个 Token 的速度编辑文件](https://mp.weixin.qq.com/s/GdVeWbxwRVu6iSTvk6NOgw) |
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.